Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-32030

SyncTail::oplogApplication uses a long-running OperationContext (3.2/3.4)

    XMLWordPrintable

    Details

    • Type: Improvement
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Fixed
    • Affects Version/s: 3.2.17, 3.4.10
    • Fix Version/s: 3.2.18, 3.4.11
    • Component/s: Replication
    • Labels:
      None
    • Backwards Compatibility:
      Fully Compatible
    • Backport Requested:
      v3.2
    • Sprint:
      Storage 2017-12-04

      Description

      Replication runs a background thread for oplogApplication that lives for the duration of the mongod process. In 3.2, that thread creates a single OperationContext and uses it for the entire lifetime.

      The same lifetime is also seen in 3.4

      I believe that if those OperationContexts are scoped to the main loop, as in master, then they would refresh frequently enough to release recovery unit resources that can impede progress of dropping collections.

        Attachments

          Issue Links

            Activity

              People

              • Votes:
                0 Vote for this issue
                Watchers:
                6 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: